Skip to content
Daytona MCP Server

Daytona MCP Server

by Daytona
Released
Repository Docs
Developer Tools

Summary

Daytona's official MCP server: give an agent a disposable cloud sandbox with a file system, git, process execution, computer use and a live preview URL.

Features

  • Create and tear down isolated cloud sandboxes on demand
  • File system and git operations inside the sandbox
  • Run processes and code, and read back their output
  • Computer use for driving a GUI in the sandbox
  • Preview URLs so a human can see what the agent just started

Installation

Set up this MCP server in your favorite AI agent — copy a ready-made configuration below.

Any MCP-compatible agent

Most agents (Claude, Cursor, Windsurf, VS Code, and more) read a standard mcpServers configuration.

  1. Open your agent's MCP configuration file.
  2. Merge the snippet below into it, filling in the environment variables with your own values.
  3. Restart the agent — the "Daytona MCP Server" tools will be available.
{
  "mcpServers": {
    "daytona-mcp": {
      "command": "daytona",
      "args": [
        "mcp",
        "start"
      ],
      "env": {
        "HOME": "${HOME}",
        "PATH": "${HOME}:/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin"
      }
    }
  }
}

Description

Daytona's official MCP server hands an AI agent its own isolated cloud sandbox — a real machine to run code on, rather than a description of one.

The tools cluster into six areas: sandbox management (create, list and tear down environments), file system operations (read, write and move files inside the sandbox), git operations (clone, branch, commit), process and code execution (run commands and scripts and read their output), computer use (drive a GUI inside the sandbox), and preview (get a public URL for a service the agent has just started, so a human can look at what was built).

The distinction that matters is where the code runs. Executing agent-written code on the developer's own laptop means the agent inherits the developer's credentials, SSH keys and file system. A Daytona sandbox is created for the task, has only what you put in it, and is thrown away afterwards — which makes it a reasonable place to run something you have not read yet.

Setup

Install the CLI (brew install daytonaio/cli/daytona on macOS and Linux, or irm https://get.daytona.io/windows | iex on Windows), authenticate with daytona login, then wire up your client with daytona mcp init claude, daytona mcp init cursor or daytona mcp init windsurf. daytona mcp config prints the JSON block if you would rather paste it yourself. A Daytona account is required.

Related MCP Servers

MCP: Replicate MCP Server

by Replicate

New

Replicate's official MCP server: search thousands of hosted models, read their schemas, and run predictions on image, video, audio and language models from inside an agent.

Developer Tools

MCP: Hugging Face

by Hugging Face

New

Hugging Face's official remote MCP server — search and inspect models, datasets and Spaces, browse the Hub filesystem over hf:// URIs, and turn Gradio Spaces into callable tools.

Developer Tools

MCP: Resend

by Resend

New

Resend's official MCP server for transactional and marketing email — send, schedule and batch messages, manage templates, contacts, broadcasts, automations, domains and suppressions from an agent.

Developer Tools
2 views

MCP: Postman MCP Server

by Postman

New

Postman's official MCP server — reach your workspaces, collections, specs and environments from Claude Code, Cursor, VS Code or Copilot, in four tool tiers.

Developer Tools
4 views
Browse all MCP servers →